Research progress and prospects of atmospheric motion vector based on meteorological satellite images
-
摘要: 本文主要回顾了气象卫星导风的发展历程并对未来发展方向进行了阐述. 引言部分首先回顾发展历程并介绍了导风发展史上的一些里程碑事件,分别对中国、美国、欧洲以及日本的气象卫星导风情况进行了简要介绍. 第一节详细总结了多种传统气象卫星导风算法的特点和关键技术,介绍了交叉相关法、形态辨认法以及亚像元法. 此外,还描述了五种较为常用的高度指定算法,对传统导风追踪算法以及高度指定算法的原理进行了详细地描述. 第二节归纳了最近几年基于计算机视觉和机器学习技术发展起来的多种新体制的气象卫星导风产品,分别介绍了光流法、三维导风以及中尺度导风的优势和研究背景. 最后,对比了传统与新型气象卫星各种导风算法的优缺点,展望了未来的发展趋势和应用前景. 特别是指出光流法导风有着空间分辨率高、三维导风能得到更多层风场的信息、中尺度导风则能对特殊天气如热带气旋实现高时空分辨率的观测的优势,并提出三维导风与中尺度导风的应用研究将是未来重要发展方向.Abstract: This paper mainly reviews the history and prospects of the atmospheric motion vector (AMV) of meteorological satellites. The development history of AMV and some milestone events are first introduced before briefly discussing them in the contexts of China, the United States, Europe, and Japan. The first section provides a detailed summary of the characteristics and key technologies of various traditional AMV algorithms, introduces the cross-correlation, pattern recognition, and nested tracking approaches, and describes five commonly used height assignment algorithms and their basic principles. The second section discusses several recently developed AMV products based on computer vision and machine learning technologies and introduces the advantages and research histories of the optical flow method, and three-dimensional and mesoscale AMVs. Finally, we compare the advantages and disadvantages of new and traditional AMV algorithms before examining the potential for future applications and development trends. We specifically highlight the higher spatial resolution obtained by the advanced optical flow method, better wind field information from three-dimensional AMV, and finer spatial and temporal resolutions of special weather from mesoscale AMV such as cyclones. Furthermore, we predict more promising three-dimensional and mesoscale AMVs in the upcoming future.
-
Key words:
- atmospheric motion vector /
- three-dimensional winds /
- tracking method /
- height assignment
-
图 1 我国风云系列静止气象卫星大气导风产品的发展历程. 图片右上角数字为图像通道波长;红、绿、蓝三色分别代表高、中、低层的导风矢量
Figure 1. Development history of the AMV product of China's Fengyun geostationary series meteorological satellites. Number in the upper right corner of each picture indicates wavelength (μm); red, green, and blue vectors represent the wind vectors at the high, middle, and low layers of the atmosphere, respectively
图 3 亚像元法示意图(修改自Bresky et al., 2012)
Figure 3. Schematic diagram of nested tracking method (modified from Bresky et al., 2012)
图 4 基于(a)交叉相关法与(b)光流法反演的风场(韩雷等,2008)
Figure 4. Wind fields retrieved by (a) cross-correlation and (b) optical flow methods (Han et al., 2008)
图 6 2002年美国堪萨斯州东北部上空的中尺度导风. 绿色、蓝色、紫色分别表示1000~700 hPa、700~400 hPa、400~100 hPa层内的矢量导风. 蓝色箭头突出了成熟对流附近的中对流层分流(Bedka and Mecikalski, 2005)
Figure 6. Mesoscale AMVs in northeast Kansas in 2002; green, blue, and purple AMV barbs represent AMVs at 1000-700 hPa, 700-400 hPa, and 400-100 hPa layers, respectively. Blue arrows highlight mid-tropospheric diffluence in the vicinity of the mature convection (Bedka and Mecikalski, 2005)
表 1 气象卫星导风反演方法对比
Table 1. Comparison of traditional and new AMV algorithms
方法 优点 缺点 传统导风方法 CC法 实现简单,精确,
空间分辨率较低耗费算力较大 模型识别法 适用范围广 精度不够高 亚像元法 降低慢速偏差,提高空间分辨率 耗费算力较大 新型导风方法 光流法 计算速度快,得出风场较平滑,能追踪风速大的矢量,分辨率高 耗费算力较大 三维导风 能得到多层导风,数据信息量巨大 时间频次较低,对红外高光谱探测器定标精度要求较高 中尺度导风 能够得到高时空分辨率导风产品 空间覆盖范围有限 -
[1] Abidi M A, Gonzalez R C. 1987. Cloud motion measurement from radar image sequences[C]// Cambridge Symposium. International Society for Optics and Photonics. SPIE, 846: 54-60. [2] Adelson E H, Anderson C H, Bergen J R, et al. 1984. Pyramid methods in image processing[J]. RCA Engineer, 29(6): 33-41. [3] Apke J M, Mecikalski J R, Bedka K, et al. 2018. Relationships between deep convection updraft characteristics and satellite-based super rapid scan mesoscale atmospheric motion vector–derived flow[J]. Monthly Weather Review, 146(10): 3461-3480. doi: 10.1175/MWR-D-18-0119.1 [4] Arking A, Lo R C, Rosenfeld A. 1978. A Fourier approach to cloud motion estimation[J]. Journal of Applied Meteorology and Climatology, 17(6): 735-744. doi: 10.1175/1520-0450(1978)017<0735:AFATCM>2.0.CO;2 [5] Bedka K M, Mecikalski J R. 2005. Application of satellite-derived atmospheric motion vectors for estimating mesoscale flows[J]. Journal of Applied Meteorology and Climatology, 44(11): 1761-1772. doi: 10.1175/JAM2264.1 [6] Borde R, Oyama R. 2008. A direct link between feature tracking and height assignment of operational atmospheric motion vectors[C]// Ninth International Winds Workshop. [7] Borde R, Doutriaux-Boucher M, Dew G, et al. 2014. A direct link between feature tracking and height assignment of operational EUMETSAT atmospheric motion vectors[J]. Journal of Atmospheric and Oceanic Technology, 31(1): 33-46. doi: 10.1175/JTECH-D-13-00126.1 [8] Borde R, García-Pereda J. 2014. Impact of wind guess on the tracking of atmospheric motion vectors[J]. Journal of Atmospheric and Oceanic Technology, 31(2): 458-467. doi: 10.1175/JTECH-D-13-00105.1 [9] Bormann N, Thépaut J N. 2004. Impact of MODIS polar winds in ECMWF’s 4DVAR data assimilation system[J]. Monthly Weather Review, 132(4): 929-940. doi: 10.1175/1520-0493(2004)132<0929:IOMPWI>2.0.CO;2 [10] Bresky W, Daniels J. 2006. The feasibility of an optical flow algorithm for estimating atmospheric motion[C]//Proceedings of the Eighth Int. Winds Workshop, Beijing, China, 24-28. [11] Bresky W C, Daniels J M, Bailey A, et al. 2012. New methods toward minimizing the slow speed bias associated with atmospheric motion vectors[J]. Journal of Applied Meteorology and Climatology, 51(12): 2137-2151. doi: 10.1175/JAMC-D-11-0234.1 [12] Büche G, Karbstein H, Kummer A, et al. 2006. Water vapor structure displacements from cloud-free Meteosat scenes and their interpretation for the wind field[J]. Journal of Applied Meteorology and Climatology, 45(4): 556-575. doi: 10.1175/JAM2343.1 [13] Campbell G, Holmlund K. 2004. Geometric cloud heights from Meteosat[J]. International Journal of Remote Sensing, 25(21): 4505-4519. doi: 10.1080/01431160410001726076 [14] Carr J L, Wu D L, Wolfe R E, et al. 2019. Joint 3D-wind retrievals with stereoscopic views from MODIS and GOES[J]. Remote Sensing, 11(18): 2100. doi: 10.3390/rs11182100 [15] Carr J L, Wu D L, Daniels J, et al. 2020. GEO–GEO stereo-tracking of Atmospheric Motion Vectors (AMVs) from the geostationary ring[J]. Remote Sensing, 12(22): 3779. doi: 10.3390/rs12223779 [16] Dew G, Holmlund K. 2000. Investigations of cross-correlation and euclidian distance target matching techniques in the MPEF environment[C]//Fifth International Winds Workshop. [17] Dixon M, Wiener G. 1993. TITAN: Thunderstorm identification, tracking, analysis, and nowcasting—A radar-based methodology[J]. Journal of Atmospheric and Oceanic Technology, 10(6): 785-797. doi: 10.1175/1520-0426(1993)010<0785:TTITAA>2.0.CO;2 [18] Endlich R M, Wolf D E, Hall D J, et al. 1971. Use of a pattern recognition technique for determining cloud motions from sequences of satellite photographs[J]. Journal of Applied Meteorology and Climatology, 10(1): 105-117. doi: 10.1175/1520-0450(1971)010<0105:UOAPRT>2.0.CO;2 [19] Endlich R M, Wolf D E. 1981. Automatic cloud Ttracking applied to GOES and METEOSAT observations[J]. Journal of Applied Meteorology and Climatology, 20(3): 309-319. doi: 10.1175/1520-0450(1981)020<0309:ACTATG>2.0.CO;2 [20] Eyre J R, English S J, Forsythe M. 2020. Assimilation of satellite data in numerical weather prediction. Part I: The early years[J]. Quarterly Journal of the Royal Meteorological Society, 146(726): 49-68. doi: 10.1002/qj.3654 [21] Farnebäck G. 2002. Polynomial expansion for orientation and motion estimation[D]. Linköping, Sweden: Linköping University. [22] Farnebäck G. 2003. Two-frame motion estimation based on polynomial Eexpansion[C]//Bigun J, Gustavsson T. Image Analysis. Berlin, Heidelberg: Springer, 363-370. [23] Fujita T. 1963. A Technique for Precise Analysis of Satellite Photographs[M]. Chicago, America: University of Chicago. [24] Fujita T. 1965. Evaluation of errors in the graphical rectification of satellite photographs[J]. Journal of Geophysical Research, 70(24): 5997-6007. doi: 10.1029/JZ070i024p05997 [25] García-Pereda J, Borde R. 2014. The impact of the tracer size and the temporal gap between images in the extraction of atmospheric motion vectors[J]. Journal of Atmospheric and Oceanic Technology, 31(8): 1761-1770. doi: 10.1175/JTECH-D-13-00235.1 [26] 韩丰, 魏鸣, 李南, 等. 2013. 反射率因子和径向速度共同约束反演多普勒雷达风场[J]. 遥感学报, 17(3): 578-589 doi: 10.11834/jrs.20132139Han F, Wei M, Li N, et al. 2013. Doppler radar wind field retrieval using reflectivity and radial velocity data[J]. Journal of Remote Sensing, 17(3): 578-589 (in Chinese). doi: 10.11834/jrs.20132139 [27] 韩雷, 王洪庆, 林隐静. 2008. 光流法在强对流天气临近预报中的应用[J]. 北京大学学报: 自然科学版, 229(5): 751-755 doi: 10.13209/j.0479-8023.2008.116Han L, Wang H Q, Lin Y J. 2008. Application of optical flow method in strong convective weather nowcasting forecast[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 229(5): 751-755 (in Chinese). DOI: 10.13209/j.0479-8023.2008.116. [28] Hayden C M, Purser R J. 1995. Recursive filter objective analysis of meteorological fields: Applications to NESDIS operational processing[J]. Journal of Applied Meteorology and Climatology, 34(1): 3-15. doi: 10.1175/1520-0450-34.1.3 [29] Héas P, Mémin É. 2008. Optical-flow for 3D atmospheric motion estimation. [C]// VISAPP 2008 - 3rd International Conference on Computer Vision Theory and Applications. [30] Holmlund K. 1998. The utilization of statistical properties of satellite-derived atmospheric motion vectors to derive quality indicators[J]. Weather and Forecasting, 13(4): 1093-1104. doi: 10.1175/1520-0434(1998)013<1093:TUOSPO>2.0.CO;2 [31] Holmlund K, Velden C S, Rohn M. 2001. Enhanced automated quality control applied to high-density satellite-derived winds[J]. Monthly Weather Review, 129(3): 517-529. doi: 10.1175/1520-0493(2001)129<0517:EAQCAT>2.0.CO;2 [32] 贾静荣, 孙逢林, 邱启荣, 等. 2018. 基于修正TV范数光流法的Himawari-8云导风算法研究[J]. 大气与环境光学学报, 13(6): 453-461Jia J R, Sun F L, Qiu Q R, et al. 2018. Research on cloud derived wind algorithm of Himawari-8 based on modified TV norm optical flow method[J]. Journal of Atmospheric and Environmental Optics, 13(6): 453-461(in Chinese). [33] Johnson J T, MacKeen P L, Witt A, et al. 1998. The storm cell identification and tracking algorithm: An enhanced WSR-88D algorithm[J]. Weather and Forecasting, 13(2): 263-276. doi: 10.1175/1520-0434(1998)013<0263:TSCIAT>2.0.CO;2 [34] Jung J, Marshall J L, Daniels J, et al. 2010. Investigating height assignment type errors in the NCEP global forecast system[C]// Proceedings of 10th International Wind Workshop, Tokyo, Session 3, Paper 4. [35] Kodaira N, Kato K, Hamada T. 1979. Man-machine interactive processing of extracting cloud top height and cloud wind data from the GMS images[J]. Meteorological Satellite Center Technical Note, 1: 59-78. [36] Lazzara M A, Dworak R, Santek D A, et al. 2014. High-latitude atmospheric motion vectors from composite satellite data[J]. Journal of Applied Meteorology and Climatology, 53(2): 534-547. doi: 10.1175/JAMC-D-13-0160.1 [37] Leese J A, Novak C S, Clark B B. 1971. An automated technique for obtaining cloud motion from geosynchronous satellite data using cross correlation[J]. Journal of Applied Meteorology and Climatology, 10(1): 118-132. doi: 10.1175/1520-0450(1971)010<0118:AATFOC>2.0.CO;2 [38] Lim A H N, Nebuda S E, Jung J A, et al. 2022.Optimizing the assimilation of the GOES-16/-17 atmospheric motion vectors in the hurricane weather forecasting (HWRF) model[J]. Remote Sensing, 14: 3068. https://doi.org/10.3390/rs14133068. [39] 林文明, 郎姝燕, 赵晓康, 等. 2021. 中法海洋卫星散射计近海岸海面风场反演研究[J]. 海洋学报, 43(10): 115-123Lin W M, Lang S Y, Zhao X K, et al. 2021. Inversion of near-shore sea surface wind field of China-France Marine satellite Scatterometer[J]. Acta Oceanologica Sinica, 43(10): 115-123 (in Chinese). [40] Ma Z, Li J, Han W, et al. 2021. Four-dimensional wind fields from geostationary hyperspectral infrared sounder radiance measurements with high temporal resolution[J]. Geophysical Research Letters, 48(14): e2021GL093794. [41] 马铮. 2022. 静止气象卫星大气探测的反演与短临预报应用研究[D]. 北京: 中国科学院大气物理所.Ma Z. 2022. Studies of the retrieval and nowcasting applications of hyperspectral infrared sounders onboard the geostationary meteorological satellites[D]. Beijing: Institute of Atmospheric Physics, Chinese Academy of Sciences (in Chinese). [42] Marshall J L, Pescod N, Seaman B, et al. 1994. An operational system for generating cloud drift winds in the Australian region and their impact on numerical weather prediction[J]. Weather and Forecasting, 9(3): 361-370. doi: 10.1175/1520-0434(1994)009<0361:AOSFGC>2.0.CO;2 [43] Maschhoff K R, Polizotti J J, Aumann H H, et al. 2016. Mistic winds, a microsatellite constellation approach to high-resolution observations of the atmosphere using infrared sounding and 3D winds measurements[J]. Sensors, Systems, and Next-Generation Satellites XX. SPIE, 10000: 85-98. [44] Mecikalski J R, Bedka K M. 2006. Forecasting convective initiation by monitoring the evolution of moving cumulus in daytime GOES imagery[J]. Monthly Weather Review, 134: 49-78. [45] Menzel W P, Stewart T R, Smith W L. 1983. Improved cloud motion wind vector and altitude assignment using VAS[J]. Journal of Applied Meteorology, 22(3): 377-384. [46] Menzel W P. 2001. Cloud tracking with satellite imagery: From the pioneering work of Ted Fujita to the present[J]. Bulletin of the American Meteorological Society, 82(1): 33-48. doi: 10.1175/1520-0477(2001)082<0033:CTWSIF>2.3.CO;2 [47] Min M, Wu C, Li C, et al. 2017. Developing the science product algorithm testbed for Chinese next-generation geostationary meteorological satellites: Fengyun-4 series[J]. Journal of Meteorological Research, 31(4): 708-719. doi: 10.1007/s13351-017-6161-z [48] Nieman S J, Menzei W P, Hayden C M, et al. 1997. Fully automated cloud-drift winds in NESDIS operations[J]. Bulletin of the American Meteorological Society, 78(6): 1121-1134. doi: 10.1175/1520-0477(1997)078<1121:FACDWI>2.0.CO;2 [49] Novak C, Young M. 1977. The operational processing of wind estimates from cloud motions: Past, present and future[C]//ERIM Proceedings of the 11th International Symposium on Remote Sensing of Environment, Vol. 2. [50] Santek D, Dworak R, Nebuda S, et al. 2019a. 2018 Atmospheric Motion Vector (AMV) intercomparison study[J]. Remote Sensing, 11(19): 2240. doi: 10.3390/rs11192240 [51] Santek D, Nebuda S, Stettner D. 2019b. Demonstration and evaluation of 3D winds generated by tracking features in moisture and ozone fields derived from AIRS sounding retrievals[J]. Remote Sensing, 11(22): 2597. doi: 10.3390/rs11222597 [52] Shenk W E. 1991. Suggestions for improving the derivation of winds from geosynchronous satellites[J]. Global and Planetary Change, 4(1): 165-171. [53] Smith E A, Phillips D R. 1972. Automated cloud tracking using precisely aligned digital ATS pictures[J]. IEEE Transactions on Computers, C-21(7): 715-729. doi: 10.1109/T-C.1972.223574 [54] Stoffelen A, Marseille G J, Bouttier F, et al. 2006. ADM-aeolus doppler wind lidar observing system simulation experiment[J]. Quarterly Journal of the Royal Meteorological Society, 132(619): 1927-1947. doi: 10.1256/qj.05.83 [55] 隋新秀, 王振会, 鲍艳松, 等. 2018. FY-2E晴空风矢同化对台风分析和预报的影响研究[J]. 热带气象学报, 34(6): 819-831 doi: 10.16032/j.issn.1004-4965.2018.06.010Sui X X, Wang Z H, Bao Y S, et al. 2018. Influence of FY-2E clear sky vector assimilation on typhoon analysis and forecast[J]. Journal of Tropical Meteorology, 34(6): 819-831 (in Chinese). doi: 10.16032/j.issn.1004-4965.2018.06.010 [56] Sun F, Min M, Qin D, et al. 2019a. Refined typhoon geometric center derived from a high spatiotemporal resolution geostationary satellite imaging system[J]. IEEE Geoscience and Remote Sensing Letters, 16(4): 499-503. doi: 10.1109/LGRS.2018.2876895 [57] Sun F, Qin D, Min M, et al. 2019b. Convective initiation nowcasting over China from Fengyun-4A measurements based on TV-L1 optical flow and BP_adaboost neural network algorithms[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(11): 4284-4296. doi: 10.1109/JSTARS.2019.2952976 [58] Szejwach G. 1982. Determination of semi-transparent cirrus cloud temperature from infrared radiances: Application to METEOSAT[J]. Journal of Applied Meteorology and Climatology, 21(3): 384-393. doi: 10.1175/1520-0450(1982)021<0384:DOSTCC>2.0.CO;2 [59] Velden C S, Hayden C M, Nieman S J W, et al. 1997. Upper-tropospheric winds derived from geostationary satellite water vapor observations[J]. Bulletin of the American Meteorological Society, 78(2): 173-196. doi: 10.1175/1520-0477(1997)078<0173:UTWDFG>2.0.CO;2 [60] Velden C, Dengel G, Dengel R, et al. 2004. Determination of wind vectors by tracking features on sequential moisture analyses derived from hyperspectral IR satellite soundings[C]// 13th Annual American Meteorological Society Conference on Satellite Meteorology and Oceanography. [61] Velden C, Daniels J, Stettner D, et al. 2005. Recent innovations in Dderiving tropospheric winds from meteorological satellites[J]. Bulletin of the American Meteorological Society, 86: 205-223. doi: 10.1175/BAMS-86-2-205 [62] Velden C S, Bedka K M. 2009. Identifying the uncertainty in determining satellite-derived atmospheric motion vector height attribution[J]. Journal of Applied Meteorology and Climatology, 48(3): 450-463. doi: 10.1175/2008JAMC1957.1 [63] Walker J R, MacKenzie W M, Mecikalski J R, et al. 2012. An enhanced geostationary satellite–based convective initiation algorithm for 0-2-h nowcasting with object tracking[J]. Journal of Applied Meteorology and Climatology, 51(11): 1931-1949. doi: 10.1175/JAMC-D-11-0246.1 [64] Wu T C, Liu H, Majumdar S J, et al. 2014. Influence of assimilating satellite-derived atmospheric motion vector observations on numerical analyses and forecasts of tropical cyclone track and intensity[J]. Monthly Weather Review, 142(1): 49-71. doi: 10.1175/MWR-D-13-00023.1 [65] Xie Y, Chen M, Zhang S, et al. 2022. Impacts of FY-4A atmospheric motion vectors on the Henan 7.20 rainstorm forecast in 2021[J]. Remote Sensing, 14(22): 5637. doi: 10.3390/rs14225637 [66] Xu J, Holmlund K, Zhang Q, et al. 2002. Comparison of two schemes for derivation of atmospheric motion vectors[J]. Journal of Geophysical Research: Atmospheres, 107(D14): ACL 4-1-ACL 4-15. [67] 许健民, 张其松. 2006. 卫星风推导和应用综述[J]. 应用气象学报. 17(5): 574-582 doi: 10.3969/j.issn.1001-7313.2006.05.007Xu J M, Zhang Q S. 2006. Review on derivation and application of satellite wind[J]. Journal of Applied Meteorology Science, 17(5): 574-582 (in Chinese). doi: 10.3969/j.issn.1001-7313.2006.05.007 [68] Yang J, Zhang Z, Wei C, et al. 2017. Introducing the new generation of Chinese geostationary weather satellites, Fengyun-4[J]. Bulletin of the American Meteorological Society, 98(8): 1637-1658. doi: 10.1175/BAMS-D-16-0065.1 [69] 张其松, 许健民, 张晓虎. 2011. 风云二号卫星红外通道风高度指定改进研究[C]//第28届中国气象学会年会——S2风云卫星定量应用与数值, 180-187.Zhang Q S, Xu J M, Zhang X H. 2011. Study on improvement of wind altitude designation in infrared channel of Fengyun-2 satellite[C]//The 28th Annual Meeting of Chinese Meteorological Society-Quantitative Application and Numerical Value of S2 Fengyun Satellite, 180-187 (in Chinese). [70] 张晓虎, 张其松, 许健民. 2017. 半透明云风矢量高度算法中代表运动像元的使用[J]. 应用气象学报, 28(3): 270-282 doi: 10.11898/1001-7313.20170302Zhang X H, Zhang Q S, Xu J M. 2017. Use of representative motion pixels in semi-transparent cloud-wind vector height algorithm[J]. Journal of Applied Meteorology Science, 28(3): 270-282 (in Chinese). doi: 10.11898/1001-7313.20170302 -